C 打印最长的单词

Question

I wrote this program to find and print out the longest word (continuous string of [a-zA-Z] characters), where the string is user input. The code works but does have a bug, If I input any special characters like '!' the output will print out the longest word along with some other weird characters. If I do not enter any special characters the program works as intended. I am sure this is something simple, but I haven't been able to figure out what is causing this bug. Also any methods of improving the code would be appreciated.

Thank you all for your time.

#include <stdio.h>
#include <string.h>

#define MAX_INPUT 1000

void LongestWord(char sen[]);

int main(void) {
  char input[MAX_INPUT];

  LongestWord(fgets(input, MAX_INPUT, stdin));
  return 0;
}

void LongestWord(char sen[]) {
    int i, current_len, longest_len, input_length;

    input_length = strlen(sen);

    char temp_string[input_length];
    char longest_string[input_length];

    current_len = longest_len = 0;

    for (i = 0; i < input_length; i++) {
        //If it is a letter
        if ((sen[i] >= 97 && sen[i] <= 122) || (sen[i] >= 65 && sen[i] <= 90)) {
            temp_string[current_len] = sen[i];
            current_len++;
        } else {
            if (current_len > longest_len) {
                //Clear longest string
                memset(longest_string, 0, sizeof(longest_string));

                //Copy temp_string to longest string
                strcpy(longest_string, temp_string);

                //set longest length equal to current
                longest_len = current_len;

                //Clear temp_string
                memset(temp_string, 0, sizeof(temp_string));
            }
            current_len = 0;
        }
    }

    printf("Longest string: %s\n", longest_string);
}

Answer 1

The problems are:

• temp_string and longest_string should be defined with a size of input_length + 1`
• you should set longest_string to an empty string with *longest_string = '\\0'; to avoid printing an uninitialized array if there are no letters on the line.
• avoid hardcoding ASCII values in your code, it is not readable. You could use (sen[i] >= 'a' && sen[i] <= 'z') || (sen[i] >= 'A' && sen[i] <= 'Z') (sen[i] >= 'a' && sen[i] <= 'z') || (sen[i] >= 'A' && sen[i] <= 'Z') or better use isalpha((unsigned char)sen[i]) from <ctype.h> .
• you only test for the longest string if you find a byte that is not a letter. This prevents the last word from being included in the search. You should instead first find a letter, then match all subsequent letters and test if you have found a longer word.
• the calls to memset are overkill. You should just set the null terminator in temp_string with temp_string[current_len] = '\\0'; so strcpy can copy it as a C string and get rid of both memset calls.

You don't actually need to copy the substrings, just keep track of the offsets for the current and longest word:

#include <ctype.h>
#include <stdio.h>

void LongestWord(const char *sen) {
    int i, len, best_start, best_len;

    best_start = best_len = 0;
    for (i = 0; sen[i] != '\0'; i++) {
        if (isalpha((unsigned char)sen[i]) {
            // we have a letter: compute the word length
            for (len = 1; isalpha((unsigned char)sen[i + len]; len++)
                continue;
            if (len > best_len) {
                // remember the longest word's offset and length
                best_start = i;
                best_len = len;
            }
            // skip all letters (-1 because of i++ in the for loop)
            i += len - 1;
        }
    }
    // use `%.*s` to print a substring with printf
    printf("Longest string: %.*s\n", best_len, sen + best_start);
}